The identification of an individual by means of the record of a body imprint taken from a region of the body thereof, and in particular most commonly by means of a fingerprint taken from the distal phalanx of one of the fingers thereof is confronted with numerous difficulties.
When capturing the body imprint, one problem consists in protecting against fraud committed by the fake body imprint technique, and in particular the so-called “model finger” technique, most frequently in the case of capturing a fingerprint. Solutions are known which enable fraud of this type to be detected (see, for example, documents FR 2 881 856 and FR 06 09629 by the applicant).
When capturing the body imprint, another problem lies in the positional deviation between the region of the body bearing the body imprint and the position taken as the reference position, which was adopted during a previous registration capture (also called the enrollment step), which provided an image held in memory. Technical solutions are known for identifying the individual, which enable comparison of the captured image of the body imprint with the stored image of the body imprint, these two images not being strictly coincident. However, such processing of the data is very complex and long.
Admittedly, in order to solve this problem, the presence of one or more reference marks could be anticipated on the region of the body from which the imprint is to be taken. In this way, provisions could be made for marking said region of the body with one or more visible or invisible marks, e.g., such as two lines in predetermined mutual positions (e.g., parallel positions). However, it is not foreseeable, for example, to mark the body by making incisions in order to obtain unalterable marks, and one would have to make do with a marking in indelible ink, for example: the disadvantage of this solution lies in the limited lifetime of these marks and the application thereof would be of a small scale. It could likewise be anticipated to detect a small number of characteristic points present in the region of the body (e.g., minutiae and deltas of a fingerprint). However, such reference marks are too imprecise and furthermore would require overly complex and lengthy processing of the data in order to utilize same.
Yet another problem occurs when it is a matter of identifying an individual by comparing the captured image of the body imprint taken from a region of the body of same with the stored images of body imprints from a collection containing a large number of stored body imprint images. This selection process is long, due to the large number of data items to be processed. It is known to accelerate the process by implementing another identification criteria making it possible to carry out a faster preselection of a limited number of stored body imprint images from among which the comparison with the captured body imprint image is finally carried out (see, for example, the document FR 2 810 226 by the applicant).
It is understood from the preceding that technical arrangements are known, which enable the aforementioned problems to be resolved. However, the disadvantages of these currently known solutions lies in the fact that they are suitable for dealing separately with a specific problem, and that dealing with all of the stated problems concomitantly involves parallel implementation of the various respective individual solutions. This alternative is not satisfactory, in particular due to the multiplication of the means to be implemented, and to the additional cost and possible loss of processing time (or insufficient gain in processing time) resulting therefrom.
Moreover, the solutions known to date for dealing with the respective stated problems are admittedly worthy of existence and of providing satisfaction from a technical standpoint, as far as the results obtained are concerned, but they do not optimally meet the requirements of data processing speed and cost (implementation of a minimum number of means).
In order to attempt to remedy the aforementioned disadvantages, it is known to make further use of a capture of the underlying venous network in said region of the individual's body. From the document WO 2006/043252, in particular, a system is known for identifying an individual by recording a body imprint, in particular a fingerprint, from a region of the body of same, this system comprising means of optically capturing an image of a body imprint, in particular a fingerprint, from said region of the individual's body, means of optically capturing an image of at least a portion of the venous network in said region of the individual's body and underlying said body imprint, said system comprising two lighting means capable of emitting two light beams, respectively, containing two light radiations having two separate wavelengths, respectively, and sensing means capable of capturing the two images in connection with the respective wavelengths.
However, this document does not provide any information relating to the structure of such a system, while, in actual practice, systems are required which are as simple and compact as possible and the cost of which is as low as possible.
Furthermore, the aforesaid known system carries out a dual biometric control with both the capturing of the body imprint and with the capturing of the underlying venous network. Such being the case, one difficultly which arises when capturing a body imprint is that, when capturing with a view to identifying the individual, the body region (e.g., the end of a finger) is not in exactly the same position as during the initial enrollment capture for the individual. Therefore, it is necessary to reposition the captured image in relation to the stored image. The document WO 2006/043252 does not state this problem and consequently does not propose any solution.